New HTML rendering: srh-insert-document

New HTML rendering: srh-insert-document

[Lars Magne Ingebrigtsen]

I started hacking the new HTML rendering stuff today, and it's now
minimally useful.  That is, it displays images and links and text, and
doesn't do much more than that.

So it's not actually meant to actually work on complicated HTML.  (Where
"complicated" is defined as "has tables".)

But I made it the default, anyway, if you have an Emacs with
libxml-parse-html-region in its build, which is only the case if you
have Emacs 24 from bzr from, oh, 12 hours ago or less.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: New HTML rendering: srh-insert-document

[Lars Magne Ingebrigtsen]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

> So it's not actually meant to actually work on complicated HTML.  (Where
> "complicated" is defined as "has tables".)

Which it's why it's called "shr", as in "shur, that like rilly
complicated and stuff".

I still have no mental model of how table rendering (what with all those
colspans and rowspans and width="30%" and nested tables and and and) is
supposed to be handled, so I guess I should just start typing away and
see what I end up with.

It's like test-driven development, only without the testing.  I'm like
agile, fer shur.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: New HTML rendering: srh-insert-document

[Tassilo Horn]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

Hi!

> It's like test-driven development, only without the testing.  I'm like
> agile, fer shur.

It's totally test-driven!  Just for efficiency, we split the work.  You
do the development and let users do the testing.  Microsoft uses that
model for at least 10 years, so it cannot be wrong!

Bye,
Tassilo

Re: New HTML rendering: srh-insert-document

[Lars Magne Ingebrigtsen]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

> I still have no mental model of how table rendering (what with all those
> colspans and rowspans and width="30%" and nested tables and and and) is
> supposed to be handled, so I guess I should just start typing away and
> see what I end up with.

I went to a fab concert and had a few beers, and now I know exactly how
to implement this.  It should be a breeze.

*hickup*

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: New HTML rendering: shr-insert-document

[Lars Magne Ingebrigtsen]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

> I went to a fab concert and had a few beers, and now I know exactly how
> to implement this.  It should be a breeze.

It was!

And it works with my only test case, so I guess it's perfect now.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: New HTML rendering: shr-insert-document

[Lars Magne Ingebrigtsen]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

> And it works with my only test case, so I guess it's perfect now.

I had a peek at a few Gwene groups, and it seems to mostly do the right
thing.

It doesn't implement rowspan and colspan yet, which I guess it should.
And images in tables don't work at all, which should definitely work.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: New HTML rendering: srh-insert-document

[Steinar Bang]

>>>>> Lars Magne Ingebrigtsen <larsi@gnus.org>:

> But I made it the default, anyway, if you have an Emacs with
> libxml-parse-html-region in its build, which is only the case if you
> have Emacs 24 from bzr from, oh, 12 hours ago or less.

Out of curiousity: how did you interface with libxml2?  I assume the
return value from that function is an elisp list (of list (of list (of
...)))?  Did you use the SAX-ish API to avoid having the parsed document
in memory in two versions?  Did you use the new XmlTextReader API?

Or did you just use the good old DOM-ish API?  It's pretty fast for
something that builds a parse tree in memory.

Re: New HTML rendering: srh-insert-document

[Lars Magne Ingebrigtsen]

Steinar Bang <sb@dod.no> writes:

> Or did you just use the good old DOM-ish API?  It's pretty fast for
> something that builds a parse tree in memory.

I use the DOM-ish API.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: New HTML rendering: srh-insert-document

[Steinar Bang]

>>>>> Lars Magne Ingebrigtsen <larsi@gnus.org>:

> Steinar Bang <sb@dod.no> writes:
>> Or did you just use the good old DOM-ish API?  It's pretty fast for
>> something that builds a parse tree in memory.

> I use the DOM-ish API.

Hm...